The proposed research will examine the long-term retention of paired-associates by four- and seven-year old children. The question of development differences in memory has been attacked in many previous studies. Unfortunately, only a few of such studies have been methodologically sound. By far the most common methodological deficiency has been the failure to take into account variations in amount of original learning when retention differences are examined, i.e., degree of original learning has been confounded with rate of forgetting. Moreover, in those few studies that were methodologically sound, none examined age differences with retention intervals of more than a few minutes and with children ranging down to the preschool years. That such an experiment may yield important results concerning human memory development is suggested by a recent review by Campbell and Spear (1972) of theory and of relevant animal research. The proposed research is designed with these considerations in mind. The Ss will be 60 preschoolers (mean age about 4 years, 2 months) and 60 second graders (mean age about 7 years, 6 months) enrolled in schools in Evanston. At each age level the Ss will be randomly assigned into one of three retention interval (RI) conditions: 10 minutes, 48 hours, or 10 weeks. Initial paired associate (PA) learning will ensure a high degree of learning for all Ss and one that is approximately equal for the two age levels. PA retention testing will be carried out after the intervals specified above. Major interest will be in main effects of age on retention test performance and, particularly, in the interaction between age and RI.